Method of vernier digital-to-analog conversion

Abstract

A digital-to-analog conversion, including: converting signal Y using word X=M+α−αN having length Ψ=α+β digits, where M is high order digits of α long control word X, α−αN is low order digits of β long control word X, wherein α≈β; subjecting analog signal Z to three conversions, wherein, in the first conversion, signal Z1 is proportional to Mα long high order digits of X, and to reference signal Y1, where Z1=Y1×M, in the second and third conversions, signals Z2 and Z3 are proportional to Nβ long low order digits of X and to signals Y1 and Y2, respectively, where Z2=Y1×N, and Z3=Y2×N, wherein, before the conversions, α−αN low order digits of X are multiplied by αα; and adding Z1, Z2, Z3 to generate output signal Z0, wherein Y1 and Y2 relate by Y2=Y1(1±α−α), wherein α is the base of the numbering system, α is the number of digits, by which α−αN is shifted.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

The present application is the U.S. National Phase of PCT/RU2020/050138, filed on 30 Jun. 2020, which claims priority to Russian Patent Application No. 2019120873, filed on 2 Jul. 2019, the entire contents of which are incorporated herein by reference.

BACKGROUND

Field

The invention relates to the field of electrical signals conversion, in particular, to converting digital information into an analog signal.

Related Art

Two main digital-to-analog conversion methods are known:

one involves summing unit reference signals, and the other one sums those reference signals which have different weights.

The first method makes use of only one reference value weighing only one quantum when generating the output analog signal and using binary control code (see [1. Tietze U., Schenk K. Semiconductor Circuit Engineering: Reference Manual. Translated from German. M. Mir, 1982.-512 pp., page 444, FIG. 24.1, and page 245 FIG. 24.2]).

The second method makes use of reference signals with weights that depend on a digit number, and only those reference signals are summed, for which there is a “one” in the respective input code digit when binary control code is used, or a significant non-zero digit in a non-binary control code, e.g., from 1 to 9 in a decade control code (see [2. Tietze U., Schenk K. Semiconductor Circuit Engineering: Reference Manual. Translated from German. M. Mir, 1982.-512 pp., page 446, FIG. 24.3, page 446 FIG. 24.5]).

Above-mentioned digital-to-analog converters generate reference values corresponding to values of digits of the input control code X from reference signal Y, which values are summed and form discrete values of output analog value Z.

A principle underlying the method of digital-to-analog conversion may be presented as shown in FIG. 1.

Consider a parallel digital-to-analog converter (DAC) shown in FIG. 1, which has an input for a single analog (reference)—current or voltage—signal Y, digital control inputs X clocked in time with interval t, and analog signal—current or voltage—Z output, wherein the reference signal Y and output signal Z (input current or voltage and output current or voltage) may be either identical (input current or voltage and output current or voltage) or different (input voltage or current and output current or voltage).

Conversion of digital control code X in output analog signal Z is performed in one or more clock periods t, i.e., in an ultimate case, in one clock period all digits of control word X are converted into the analog value Z. Therefore the DAC_0 which implements such conversion method is referred to as parallel.

Consider the digital control code to be a position code, and the type of numeral system with base a is not important—it may be binary (a=2), decimal (a=10), hexadecimal (a=16), or any other type.

For example, a number (word) X normalized to “one” in digital binary code is:0≤X=Σ_k=0^Ψ-1q_k×α^−k<1,

where α=2; q_k∈[0,1]; ψ is the number of bits in the word X.

For a binary DAC, the output signal is determined as follows:Z=hY(α_k-12^k-1+α_k-22^k-2+ . . . +α₀2⁰),

where Y is the reference signal; q∈(α₀, α₁, . . . , α_k-1)≡[0, 1] are bit coefficients which may take the values “0” or “1”;

k=0, 1, . . . ψ−1 is binary digit number; and

h is proportionality factor.

In the case of a decimal digital control code, the number X normalized to “one” is:0≤X=Σ_k=0^Ψ-1q_k×α^−k<1,

where α=10; q_k∈[0, 1, . . . , 9]; ψ is the number of decimal digits in the word X.

Digital-to-analog converters generate reference signals which correspond to the values of the digits of the input control code X from reference signal Y, the generated signals being summed and forming discrete values of output analog value Z.

Internal contents of DAC_0 and its operation algorithm (internal operations) are not of fundamental importance for further discussion.

Analog signal Z is generated at the output of DAC_0 from analog reference signal Y input therein by means of controlling the DAC_0 with a ψ digits long digital code X the reference signal Y being the base of analog dimensional scale. Step size (sampling interval) at the output of such DAC_0 iss=Y/(α^ψ−1).

In this case Z=Y×X and “classical” conversion X⇒Z is used, wherein conversion accuracy is a linear function of accuracy of elementary electronics of DAC_0, i.e., resistor (condenser) arrays, current (voltage) switches etc., i.e., is defined by production technology of a given DAC_0 microchip and ultimate accuracy (dispersion) of nominal values of its internal circuitry.

In the conversion discussed above, both the first and the second method are prone to errors stipulated by the manufacture of DAC_0 circuitry. Main circuitry-related error factors are:

• • technology-related dispersion of parameters (precision of manufacture);
  • environmental changes (mainly temperature) effect—thermal stability;
  • changing parameters over time (ageing); and
  • external and internal noise and interference effects.

Thus, conversion performance of real DAC_0 differs from ideal ones in the shape, interval values and their position relative to coordinate axes.

To improve conversion accuracy, internal structure of DAC_0 may be made more complex, accuracy and stability of parameters of the employed elements may be improved, but this is only possible within the manufacturability limits.

The above-mentioned reasons and, ultimately, technological manufacturability, restrict the maximum possible accuracy of parallel DAC_0, which presently does not exceed 20 . . . 22 binary digits (bits) of control code X.

A digital-to-analog conversion method is known, implementation example of which is shown in FIG. 2, where an initial DAC_0 shown in FIG. 1 is used (see [3. Tietze U., Schenk K. Semiconductor Circuit Engineering: Reference Manual. Translated from German. M. Mir, 1982.-512 pp., page 447, FIG. 24.7]).

FIG. 2 shows: DAC_2, signal adder 3, DAC_4, reference signal Y and clock signal t are provided to respective inputs of DAC_2 and DAC_4, and a digits long control words M and β digits long control words a^−αN are provided to their input buses. Output of DAC_4 is connected to an input of attenuator 5 having its output connected to the first input of signal adder 3, which has its second input connected to the output of DAC_2. Resulting analog signal Z₀ is provided at the output of signal adder 3.

In this conversion method, ψ digits long digital code word X=M+a^−αN is divided into two control codes M and a^−αN (preferably, but not necessarily having half length, ψ/2): ψ=α+β, where α is the number of high-order bits of control code X, and β is the number of the low-order bits of control code X Both DACs are clocked at the same times t.

Like the aforementioned analogues, this conversion also uses a single reference signal (dimensional scale) Y.

β long low-order digits a^−αN of control code X are multiplied in digital multiplier 1 by a^α times, thus providing a β digits control word N.

First DAC_2 converts the α digits long control word Minto analog signal Z₁, second DAC_4 converts the β digits long control code N into analog signal Z₂.

Step size (sampling interval) at the output of DAC_2 is s₁=Y/α^∝>>Y/α^ψ, and step size (sampling interval) at the of DAC_4 is s₂=Y/α^β>>Y/α^ψ. In the case α≈β, steps sizes s₁≈s₂.

As such, variations of values (range) of the output signal Z₁ when control code M normalized to “one” changes from 0 to (1−α^−α) at the output of DAC_2 will be Y(1−α^−α), and the range of output signal Z₂ when control code N normalized to “one” changes from 0 to (1−α^−β) at the output of DAC_4 will be Y(1−α^−β).

In order to reduce the step size in the output signal Z₀ to the initial value, it is necessary to reduce the range of output signal Z₂ at the output of DAC_4 by a^α times by means of the analog attenuator 5 and to add the signal Z₃=Z₂/α_α thus obtained to the output signal Z₁ of DAC_2 (which has step size s₁=Y/α^∝ in the signal adder 3.

At the output of the signal adder 3 (see FIG. 2), there will be the output signal Z₀=Z₁+Z₃, which may change from 0 to Y(1−α^−ψ) with the necessary sampling interval s=Y/α^ψ.

Accuracy of conversion of reference signal Y into analog value Z₀ will be the same as in the prior art analog as shown in FIG. 1, but accuracy requirements for DAC_4 will be reduced by a^α times.

However, in the digital-to-analog conversion method discussed above it is necessary to provide for high accuracy of DAC_2 and analog attenuator 5 arranged between the output of DAC_4 and input of the signal adder 3.

The same stringent requirements as for DAC_0 shown in FIG. 1 also apply to these two analog components of the structure shown in FIG. 2. Accuracy of DAC_2 and attenuator 5 will restrict the resulting accuracy of the whole DAC shown in FIG. 2.

It is also noted that the accuracy of analog attenuator 5 not only depends on its own accuracy but also on the output resistance of DAC_4 and input resistance of signal adder 3, which may vary within the range of operational frequencies.

Conversion method implemented in the parallel DAC structure illustrated in FIG. 2 is taken as the prototype.

SUMMARY

Technical Problem

Technical result of the invention consists in improved accuracy of conversion of analog value by means of digital processing of control code and using at least two different dimensional scales Y.

Solution to Problem

The technical result is achieved in the provided method of Vernier digital-to-analog conversion, wherein reference signal Y is converted to an analog output signal Z₀ by means of control word X=M+a^−αN having a length of ψ=α+β digits, where M is high order bits of α long control word X,

a^−αN is low order bits of β long control word X, wherein α≈β,

and two parallel conversions are carried out in respect of analog signal Z, wherein, in the first conversion, first output analog signal Z₁ is proportional to M high order bits of a long control word X, and to reference signal Y₁, i.e., Z₁=Y₁×M, in the second conversion, second output analog signal Z₂ is proportional to N low order bits of β long control word X and to reference signal Y₁, i.e., Z₂=Y1×N, wherein, before said parallel conversions, digital multiplication of a^−αN low order bits of control word X by a^α times is performed (left shift by α bits), and the converted analog signals Z₁, Z₂ are summed, characterized in that third parallel conversion is performed on analog signal Z wherein third analog output signal Z₃ is proportional to N low order bits of β long control word X and to reference signal Y₂, i.e., Z₃=Y₂×N, wherein reference signals Y₁ and Y₂ are related through the following dependency:Y₂=Y₁(1±α^−α),

where a is base of numeral system, α is a number of bits, by which control code a^−αN is shifted, after which the converted analog signal Z₁, Z₂, and Z₃ are summed to provide output signal Z₀.

As such, conversion scale is selected to be the same in parallel conversion of analog signals Z₁, Z₂, and Z₃ into output signal Z₀.

Meanwhile, if reference signal Y₂ is generated according to the expression:Y₂=Y₁(1+α^−α),

then analog of output signals Z₁, Z₂, and Z₃ is carried out by their algebraic addition according to the expression:Z₀=Z₁+Z₃−Z₂.

If reference signal Y₂ is generated according to the expression:Y₂=Y₁(1−α^−α),

then analog conversion of output signals Z₁, Z₂, and Z₃ is carried out by their algebraic addition according to the expression:Z₀=Z₁+Z₂−Z₃.

Achievement of the technical result in the claimed method by the aforementioned distinguishing features will be now explained.

The claimed method is based on the principle of Vernier conversion, i.e., using at least two dimensional scales having a fractional ratio, i.e., carrying out three conversions:X→Z₁;X→Z₂;X→Z₃;(Z₁,Z₂,Z₃)→Z₀.

In other words, there is a pair of dimensional scales (reference signals Y), dependency between which is:Y₂=Y₁(1±α^−α),

wherein the structure of number X is presented as follows (artificial partition):X=X_α+βX_α+β−1. . . X₂X₁≡Mα^−∝N

where M is a α long group of high order digits, and a^−αN is a β long group of low order digits of number X in the selected numbering system. In practice, number N is a result of multiplying a^−αN by a^αtimes (shifted by α digits to the left), i.e., digitally multiplied by a^αtimes. The following expression is then possible:

$Z_1=Y_{1}M=Y_1⨯{\sum }_{k=1}^{\alpha }{q}_{\alpha }{a}^{-\alpha };$ $Z_2=Y_{1}N=Y_1⨯{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta };$ $\begin{array}{c}{Z}_3=Y_{1}N\\ =Y_1\left(1\pm a^{-\propto }\right)\\ =Y_1⨯\left({\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\pm a^{-\propto }{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right);\end{array}$ $Z_0=Z_1\mp Z_2\pm Z_3=Y_{1}M\mp Y_{1}N\pm Y_{2}N=Y_1\left[{\sum }_{k=1}^{\alpha }{q}_{\alpha }{a}^{-\alpha }\mp {\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right]\pm Y_1\left(1\pm a^{-\alpha }\right)\left[{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right]=Y_1\left\{{\sum }_{k=1}^{\alpha }{q}_m{a}^{-\alpha }\mp {\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\pm {\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }+a^{-\alpha }{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right\}.$

The following conditions should then be met:

in the algebraic addition of signals Z₁, Z₂, and Z₃ in the output analog adder, signal Z₃ should be subtracted from signal Z₂, as should be the second reference signal in accordance with the expression:

$\begin{array}{c}{Z}_0=Z_1+Z_2-Z_3\\ =Y_1\left[{\sum }_{k=1}^{\alpha }{q}_{\alpha }{a}^{-\alpha }+{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right]-Y_1\left(1-a^{-\alpha }\right)\left[{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right]\\ =Y_1\left\{{\sum }_{k=1}^{\alpha }{q}_{\alpha }{a}^{-\alpha }+a^{-\alpha }{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right\},\end{array}$

in the algebraic addition of signals Z₁, Z₂, and Z₃ in the output analog adder, signal Z₂ should be subtracted from signal Z₃ in accordance with the expression:

$\begin{array}{c}{Z}_0=Z_1+Z_2-Z_3\\ =Y_1\left[{\sum }_{k=1}^{\alpha }{q}_{\alpha }{a}^{-\alpha }+{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right]-Y_1\left(1-a^{-\alpha }\right)\left[{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right]\\ =Y_1\left\{{\sum }_{k=1}^{\alpha }{q}_{\alpha }{a}^{-\alpha }+a^{-\alpha }{\sum }_{k=1}^{\beta }{q}_{\beta }{a}^{-\beta }\right\},\end{array}$

wherein the second reference signal Y₂=Y₁(1+α^−α).

As follows from the aforementioned expressions, these conversions bring about the same result, which provides for the necessary accuracy of conversion of digital control code X into the analog value Z₀:Z₀=Y₁{Σ_k=1^αq_αα^−α+α^−αΣ_k=1^βq_βα^−β}.

BRIEF DESCRIPTION OF THE DRAWINGS

In the drawings:

FIG. 1 shows a digital-to-analog converter (DAC) known from the prior art;

FIG. 2 shows a parallel digital-to-analog converter and method using the DAC as shown in Fig. 1, known from the prior art; and

FIG. 3 shows a Vernier digital-to-analog converter and method.

DETAILED DESCRIPTION

Structural scheme of one of the possible embodiments of Vernier digital-to-analog converter (DAC) which implements the claimed method is shown in FIG. 3, which shows:

Reference signal Y₁ source 6,

Digital multiplier 7,

First DAC_8,

Reference signal Y₂ source 9,

Second DAC_10,

Third DAC_11,

Analog adder 12.

In the digital multiplier 7, β low order digits of control word a^−αN are subjected to digital multiplication by a^α times (shift by a digits to the left). Output bus of α high order digits of control word M is connected to the respective input control bus of DAC_8, and the other input of the latter being connected to the output of reference signal Y₁ source 6. Output of the DAC_8 is connected to the respective input of analog adder 12, and the other inputs of the latter being connected to the output of DAC_10 and output of DAC_11, while β low order digits of control word N (control word a^−αN, which has been digitally multiplied by a^α times (shift by α digits to the left)) are provided to the input control bus of DAC_10 and DAC_11, the other input of DAC_10 being connected to the output of reference signal Y₁ source 6, and the respective input of DAC_11 being connected to the output of reference signal Y₂ source 9, wherein the dimensional scales (reference signals) Y₁ and Y₂ are related through the following dependency:Y₂=Y₁(1±α^−α).

Analog output signal Z₀ is obtained at the output of analog adder 12.

As a numeric example, an embodiment of Vernier conversion of a decimal number into an abstract analog parameter Z is shown.

Since the numbering system is decimal, a two-digit digital word X₍₁₀₎ is partitioned into two one-digit ones, M and N, and reference signals Y will then be: Y₁=1.0; Y₂=1.1×Y₁.

EXAMPLES

There are two illustrative variants: one for the case of M₁>N and the other one for the case of M₂<N. Let M₁=7 and M₂=3, and N have the values from 0 to 9. The steps of these calculations and conversions are summarized in Table 1 and Table 2 provided below.

TABLE 1
Y2 = 1.1 × Y1
			Z1 =	Z2 =	Z3 =	Z0 =
	M1	N	Υ1 × M	Y1 × N	Y2 × N	Z1 − Z2 + Z3
	7	0	7	0	0.0	7.0
	7	1	7	1	1.1	7.1
	7	2	7	2	2.2	7.2
	7	3	7	3	3.3	7.3
	7	4	7	4	4.4	7.4
	7	5	7	5	5.5	7.5
	7	6	7	6	6.6	7.6
	7	7	7	7	7.7	7.7
	7	8	7	8	8.8	7.8
	7	9	7	9	9.9	7.9

TABLE 2
Y2 = 1.1 × Y1
			Z1 =	Z2 =	Z3 =	Z0 =
	M2	N	Y1 × M	Y1 × N	Y2 × N	Z1 − Z2 + Z3
	3	0	3	0	0.0	3.0
	3	1	3	1	1.1	3.1
	3	2	3	2	2.2	3.2
	3	3	3	3	3.3	3.3
	3	4	3	4	4.4	3.4
	3	5	3	5	5.5	3.5
	3	6	3	6	6.6	3.6
	3	7	3	7	7.7	3.7
	3	8	3	8	8.8	3.8
	3	9	3	9	9.9	3.9

The only units of the Vernier DAC (FIG. 3), to which strict requirements for precision should apply, are output analog adder 12, reference signal Y₁ source 8, and reference signal Y₂ source 9, wherein their implementation does not involve any inextricable difficulties; moreover, they are common for (high) precision electronics.

So, e.g., in case of a twenty-digit binary input word (α=β=10) and Y₁=10 V, the necessary relative accuracy of analog adder and source is Y₁ δ_α≤2⁻²⁰≈10⁻⁶ (absolute accuracy is 9.5 μV), which is readily implementable using the current microelectronic hardware components.

In the case of Y₂=0.9×Y₁ and the numeric parameters being the same, the result will be as follows:

TABLE 3
Y2 = 0.9 × Y1
			Z1 =	Z2 =	Z3 =	Z0 =
	M	N	Y1 × M	Y1 × N	Y2 × N	Z1 + Z2 − Z3
	7	0	7	0	0.0	7.0
	7	1	7	1	0.9	7.1
	7	2	7	2	1.8	7.2
	7	3	7	3	2.7	7.3
	7	4	7	4	3.6	7.4
	7	5	7	5	4.5	7.5
	7	6	7	6	5.4	7.6
	7	7	7	7	6.3	7.7
	7	8	7	8	7.2	7.8
	7	9	7	9	8.1	7.9

TABLE 4
Y2 = 0.9 × Y1
			Z1 =	Z2 =	Z3 =	Z0 =
	M	N	Y1 × M	Y1 × N	Y2 × N	Z1 + Z2 − Z3
	3	0	3	0	0.0	3.0
	3	1	3	1	0.9	3.1
	3	2	3	2	1.8	3.2
	3	3	3	3	2.7	3.3
	3	4	3	4	3.6	3.4
	3	5	3	5	4.5	3.5
	3	6	3	6	5.4	3.6
	3	7	3	7	6.3	3.7
	3	8	3	8	7.2	3.8
	3	9	3	9	8.1	3.9

In the claimed conversion method, in the case of any ratio between numbers N and M, no loss of digits occurs while the accuracy of digital-to-analog conversion is increased by ≈a^α-1 times, as the error is significantly reduced and, hence, accuracy of parallel digital-to-analog conversion is increased without stricter requirements for DAC components manufacturing technology.

Improved conversion accuracy is stipulated by the fact that, in the case of analog addition, other conditions being equal, requirements for the accuracy of adder unit are less strict than the requirements for the accuracy of analog attenuator in the case of analog division of signal Z.

Besides, accuracy of reference signals (dimensional scales) Y₁, Y₂ and their ratio needs to be provided at one point and under direct current (under constant values of current or voltage), which is significantly simpler than doing it throughout the range of output levels of Z and operational frequencies.

The following circumstance is repeatedly emphasized: for any method of digital-to-analog conversion (double and greater integration, sigma-delta, pipelined, sequential approximation, Vernier, etc.), requirements for the accuracy of analog components are only defined by the necessary accuracy of conversion.

Weighted contribution of individual analog elements of the digital-to-analog converter which implements the claimed method into the pool of acceptable errors of the device as a whole depends on its specific circuitry implementation. And, naturally, the known rule applies:

the stricter the requirements for conversion accuracy, the stricter (at least linearly) the requirements for analog components.

Requirements for accuracy and stability of reference voltage (current) sources and analog algebraic adders of input/output voltages (currents) do not depend on the chosen conversion method, and their contribution is small.

The main contribution into the pool of errors is provided by the DAC per se (which is explicitly or implicitly included in the structure of Vernier digital-to-analog conversion) via errors in voltage/current keys and R (C) arrays.

Use of identical resistors makes it possible to significantly improve the accuracy as compared to an ordinary weighted DAC, since it is comparatively easy to make a set of precision elements with identical parameters. R-2R type DACs enable shifting, but not lifting restrictions concerning the number of digits. By virtue of laser trimming of film resistors arranged on the same substrate of a hybrid microchip, 20-22 bit accuracy of DAC may be achieved.

For this reason, relaxation of requirements for DACs in the form of reducing the required number of digits ψ=α+β of control word X while retaining the resulting accuracy of conversion is of such practical importance.

An example of configuration of the elements in the structural scheme implementing the claimed conversion method is shown.

DACs 8, 10, and 11 may be configured by the following microchips: double DAC AD5763, single DAC K594PA1, K1108PA1 or similar ones. Reference signal sources 6 and 9 may be configured by microchips LT6657 (precision voltage source) or LT3092 (precision current source).

To perform digital multiplication by α times of N low order digits of control word X (left shift by α digits), shift register microchips—universal registers KR15331R8 (SN74HC164) may be used.

Claims

1. A method of Vernier digital-to-analog conversion, wherein the method comprises: converting reference signal Y using control word X=M+α−αN having a length of ψ=α+β digits, wherein M is high order digits of α long control word X, α−αN is low order digits of β long control word X, wherein α≈β,performing two parallel conversions on an analog signal Z, wherein, in a first conversion, a first output analog signal Z1 is proportional to the M high order digits of α long control word X, and to a reference signal Y1, where Z1=Y1×M, in a second conversion, a second output analog signal Z2 is proportional to N low order digits of β long control word X and to a reference signal Y1, where Z2=Y1×N, wherein, prior to the two parallel conversions, digital multiplication of the α−αN low order digits of the control word X by ααtimes is performed;performing a third parallel conversion on the analog signal Z, in which a third output analog signal Z3 is proportional to N low order digits of β long control word X, and to a reference signal Y1, where Z3=Y2×N, wherein the reference signal Y1 and the reference signal Y2 are related through a dependency Y2=Y1(1±α−α), wherein β is a base of a numbering system, and α is a number of digits by which control code α−αN is shifted; andadding the first output analog signal Z1, the second output analog signal Z2,and the third output analog signal Z3 to generate an analog output signal Z0.

2. The method of claim 1, wherein in the two parallel conversions of the first analog signal Z1 and the second output analog signal Z2, and the parallel conversion of the third output analog signal Z3, a conversion scale is selected to be the same.

3. The method of claim 1, wherein, when the reference signal Y2 is generated according to an expression Y2=Y1(1+α−α), the first output analog signal Z1, the second output analog signal Z2, and the third output analog signal Z3 are added to generate the analog output signal Z0 according to an expression: Z0=Z1+Z3−Z2.

4. The method of claim 1, wherein when the reference signal Y2 is generated according to an expression Y2=Y1(1−α−α), the first output analog Z1, the second output analog signal Z2, and the third output analog signal Z3 are added to generate the analog output signal Z0 according to an expression: Z0=Z+Z2−Z3.